By 2050, the US population will consist of 88.5 million older adults aged ? 65 years; adults aged ?85 years are the fastest growing age group. As the population ages, the number of persons with dementia will grow rapidly. Primary care clinicians will have to play an important role in delivering dementia care. In general, clinicians may fail to incorporate best practices into geriatric dementia diagnosis and treatment because of several biases. First, they may discount downstream harms of failure to plan or prepare for declining cognition. Second, they may fail to appreciate the aggregate benefit of initiating many small improvements in the patient's life. Third, clinicians may respond to real or perceived social norms (from patients, families or other clinicians) that set expectations to behave in specific ways. Fourth, clinicians may form habits that lead them to act in a similar way to past behavior even if evidence has changed. Lastly, decision fatigue, a known behavioral phenomenon at the end of clinical shifts, may also contribute to health disparities by surfacing subconscious (implicit) bias. Coordinated, point of care clinical decision support nudges, informed by behavioral economics and social psychology and delivered via electronic health records (EHRs), may help ensure persons with Alzheimer's disease receive guideline concordant care at primary care visits. Nudges could initiate early diagnosis and treatment through default orders, ensure nonpharmacological strategies to manage behavior are the first-line of treatment through social norms, deprescribe antipsychotics that are no longer useful through an EHR prompt asking clinicians to justify refills in a free text response and made visible in the patient's medical record as an ?antipsychotic justification note?. Public commitments could be used to encourage communication with social services agencies. To have broad reach in future applications, these techniques require a partnership between government, academia and private sector stakeholders. Cooperation among federal funding agencies, clinical institutions, electronic health record vendors, pharmacy benefit managers, clinicians and patient advocacy groups, will allow new interventions to undergo careful evaluation in real-world settings and get implemented rapidly when effective. We propose a round table meeting among stakeholders to discuss how to best develop choice architectures in clinical environments to encourage guideline concordant dementia decisions in primary care for diverse elderly populations. The conference will result in a white paper that will be submitted for publication, identifying low cost behavioral science insights that can improve care for elderly persons at risk of dementia. Page 1 of 10